Pancreatic adenocarcinoma is the fourth leading cause of cancer related death in the United States. Effective therapies currently do not exist. This application is based on the hypothesis that the AGR2 gene is essential for the formation or propagation of pancreatic cancer, and proposes to test the AGR2 gene's importance in a well-characterized animal model for pancreatic adenocarcinoma. The laboratory's previous work established that the AGR2 gene is actively expressed in all pancreatic adenocarcinomas. Additional studies using adenocarcinoma cell lines established that AGR2 promotes cell growth and features that are characteristic of metastasis. Because the initial experiments were performed in cell cultures grown in vitro, validation of AGR2's significance must be established in a whole animal. Recent advances in pancreatic cancer have produced several genetically engineered mouse models that closely reflect the development and progression of human pancreatic cancer. These models serve as a valuable resource to evaluate factors that influence the biology of pancreatic cancer. The K-ras gene is mutated in over 90% of pancreatic cancers. When the mutated K-ras gene is expressed in the mouse pancreas, early premalignant lesions similar to those observed in humans result. A small proportion of these lesions will progress to frank cancer. When a second mutation involving the p53 gene is added to the same mouse model, almost all of the mice will develop adenocarcinomas, of which many will also metastasize. This application proposes to examine whether the course of pancreatic adenocarcinoma is affected after genetically deleting the AGR2 gene in the two mouse models described. If the formation or propagation of premalignant lesions or cancer is successfully compromised by AGR2's absence, it would provide strong evidence of the gene's important role and provide a novel target for therapeutic intervention. This application represents a resubmission in which we provide additional data concerning the biologic relevance of the AGR2 gene to pancreatic cancer. In addition, we also provide data supporting our ability to produce the genetically defined mouse models as previously described.